On Wednesday, September 12, 2007, at 07:41AM, "Christiaan Hofman" <[EMAIL 
PROTECTED]> wrote:
>
>On 12 Sep 2007, at 4:25 PM, Adam R. Maxwell wrote:
>
>>
>> On Sep 12, 2007, at 04:34, Christiaan Hofman wrote:
>>
>>> BTW, we may have a big change in the way we handle local files in the
>>> (near?) future. And I'm not sure if saving relative paths will be
>>> compatible with those changes.
>>
>> I was planning to store a relative path as fallback along with the
>> alias for this situation.
>>
>> Thinking about this again, it may be cleaner to make them relative to
>> the home directory instead of the .bib file.  Do you recall why they
>> are .bib relative now?
>>
>> adam
>>
>
>So you can easily move the .bib file together with the library of  
>papers. It makes that the directory structure on the different  
>systems do not need to be the same, it only matters how the papers  
>are placed relative to the file. We could also make it relative to  
>the Papers Folder, but that could break Local-Url fields that expect  
>it to be different. So I think we never allowed that because of this  
>backward compatibility. I think if we should allow something relative  
>to the Home directory we should use ~ instead of a relative path.

Okay, using a document-relative path certainly makes sense in that situation.  
The question then becomes: should we keep this behavior, or break it now?  (By 
break, I mean use home-relative instead of document-relative paths).  Opinions 
from the users?  What's easier for people to deal with?  I think Mike may be 
responsible for the original implementation, so maybe he has comments.

I like the ~ path because it gives me a fixed reference frame, and I can move 
the .bib file without breaking anything else...but since the plan is to use 
aliases, this problem will go away.  In code, I guess the file objects 
(BDSKFile) could be instantiated by the BibItem with any relative path base as 
an argument, which would then be used to construct a full path in the case of 
the alias not working.  So there's no argument either way from the code 
perspective.

adam

adam

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Bibdesk-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/bibdesk-users

Reply via email to